GrƔficos

A GramƔtica de GrƔficos

library(knitr)
library(tidyverse)
library(nycflights13)
flights %>% filter(dest=="DEN") %>%
  ggplot() + 
  geom_point(aes(x=dep_time, y=dep_delay), size=1)

EstƩticas

Geometrias

Geometria para uma variĆ”vel discreta (o nĆŗmero de observaƧƵes por grupo)

flights %>% ggplot() +
  geom_bar(aes(x=origin))

Habilidade BĆ”sica de ProgramaĆ§Ć£o: Apresentando GrĆ”ficos no relatĆ³rio final

flights %>% ggplot() +
  geom_bar(aes(x=origin))

GrƔficos com uma variƔvel contƭnua - GrƔficos de histogramas

flights <- flights %>% 
  mutate(velocidade=distance/air_time) 

flights %>% 
  ggplot() +
  geom_histogram(aes(x=velocidade))

ParĆ¢metros fixos

flights %>% 
  ggplot() +
  geom_histogram(aes(x=velocidade), binwidth=1)

flights %>% 
  ggplot() +
  geom_histogram(aes(x=velocidade), binwidth=1, colour="black", fill="orange")

GrƔficos com uma variƔvel contƭnua - GrƔficos de densidade

flights %>% 
  ggplot() +
  geom_density(aes(x=velocidade))

flights %>% 
  ggplot() +
  geom_density(aes(x=velocidade), colour="blue")

flights %>% 
  ggplot() +
  geom_density(aes(x=velocidade), colour="blue", fill="blue")

flights %>% 
  ggplot() +
  geom_density(aes(x=velocidade), colour="blue", fill="blue", alpha=0.2)

media_velocidade <- flights %>% summarize(media_velocidade=mean(velocidade,na.rm=T)) %>%
  pull(media_velocidade)
flights %>% 
  ggplot() +
  geom_density(aes(x=velocidade), colour="blue", fill="blue", alpha=0.2) +
  geom_vline(aes(xintercept = media_velocidade))

flights %>% 
  ggplot() +
  geom_density(aes(x=velocidade), colour="blue", fill="blue", alpha=0.2) +
  geom_vline(aes(xintercept = media_velocidade), linetype="dashed", colour="red")

GrƔficos com uma variƔvel contƭnua e uma variƔvel discreta

flights %>% 
  ggplot() +
  geom_histogram(aes(x=dep_time, fill=origin))

flights %>% 
  ggplot() +
  geom_histogram(aes(x=dep_time, fill=origin), 
                 position = "dodge", binwidth=200)

flights %>% 
  ggplot() +
  geom_density(aes(x=dep_time, colour=origin))

flights %>% 
  ggplot() +
  geom_density(aes(x=dep_time, fill=origin))

flights %>% 
  ggplot() +
  geom_density(aes(x=dep_time, fill=origin), alpha=0.5)

flights %>% 
  ggplot() +
  geom_density(aes(x=dep_time, colour=origin, fill=origin), alpha=0.5)

GrƔficos com uma variƔvel contƭnua e uma variƔvel discreta - GrƔficos de boxplot

flights %>% 
  ggplot() +
  geom_boxplot(aes(x=origin, y=dep_time))

GrĆ”ficos com um valor Ćŗnico por uma variĆ”vel discreta - GrĆ”ficos de coluna

flights %>% group_by(origin) %>% 
  summarize(dep_delay_media=mean(dep_delay,na.rm=T)) %>% 
  ggplot() +
  geom_col(aes(x=origin, y=dep_delay_media))

GrƔficos de duas variƔveis contƭnuas

flights %>% sample_n(1000) %>% 
  ggplot() + 
  geom_point(aes(x = distance, y = air_time))

flights %>% sample_n(1000) %>% 
  ggplot() + 
  geom_point(aes(x = distance, y = air_time), size=0.1, color="blue", shape=2)

flights %>% sample_n(1000) %>% 
  ggplot() + 
  geom_point(aes(x = distance, y =air_time), size=0.1) +
  geom_smooth(aes(x = distance, y = air_time), method = "lm", se = FALSE)

flights %>%# sample_n(50) %>% 
  ggplot() + 
  geom_point(aes(x = distance, y = air_time), size=0.1) +
  geom_smooth(aes(x = distance, y = air_time), method = "lm")

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time), size=0.1) +
  geom_smooth(aes(x = distance, y = air_time), method = "loess")

GrĆ”ficos de trĆŖs ou mais variĆ”veis

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = dep_time, y = dep_delay, size=distance))

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = dep_time, y = dep_delay, colour=origin))

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = dep_time, y = dep_delay, shape=origin))

MĆŗltiplos GrĆ”ficos (facet_grid)

flights %>% sample_n(1000) %>% 
  ggplot() + 
  geom_point(aes(x = dep_time, y = dep_delay)) + 
  facet_grid(cols=vars(origin))

flights %>% sample_n(1000) %>% 
  ggplot() + 
  geom_point(aes(x = dep_time, y = dep_delay)) + 
  facet_grid(rows=vars(origin))

flights %>% sample_n(1000) %>% 
  ggplot() + 
  geom_point(aes(x = dep_time, y = dep_delay)) + 
  facet_grid(rows=vars(month), cols=vars(origin))

GrƔficos de Linha

flights %>% 
  mutate(month=factor(month, levels=1:12, ordered=TRUE)) %>%
  group_by(month) %>%
  summarize(dep_delay_media=mean(dep_delay,na.rm=T)) %>%
  ggplot() +
  geom_line(aes(x=month, y=dep_delay_media), group=1)

flights %>% 
  mutate(month=factor(month, levels=1:12, ordered=TRUE)) %>%
  group_by(month, origin) %>%
  summarize(dep_delay_media=mean(dep_delay,na.rm=T)) %>%
  ggplot() +
  geom_line(aes(x=month, y=dep_delay_media, group=origin))

flights %>% 
  mutate(month=factor(month, levels=1:12, ordered=TRUE)) %>%
  group_by(month, origin) %>%
  summarize(dep_delay_media=mean(dep_delay,na.rm=T)) %>%
  ggplot() +
  geom_line(aes(x=month, y=dep_delay_media, group=origin, colour=origin))

GrƔfico de barras 100%

flights %>% 
  mutate(month=factor(month, levels=1:12, ordered=TRUE)) %>%
  group_by(month, origin) %>%
  summarize(dep_delay_total=sum(dep_delay,na.rm=T)) %>%
  ggplot() +
  geom_col(aes(x=month, y=dep_delay_total, fill=origin), position = "fill")

Mais geometrias

flights %>% sample_n(100) %>% 
  ggplot() + 
  geom_text(aes(x = dep_time, y = dep_delay, label=dest))

flights %>% group_by(origin, month) %>% 
  summarize(dep_delay_media=mean(dep_delay,na.rm=T)) %>% 
  ggplot() + 
  geom_tile(aes(x = origin, y = month, fill=dep_delay_media))

flights %>% group_by(origin) %>%
  summarize(dep_delay_total=sum(dep_delay, na.rm=T)) %>%
  ggplot() + 
  geom_col(aes(x="", y=dep_delay_total, fill=origin), position="fill")

flights %>% group_by(origin) %>%
  summarize(dep_delay_total=sum(dep_delay, na.rm=T)) %>%
  ggplot() + 
  geom_col(aes(x="", y=dep_delay_total, fill=origin), position="fill") +
  coord_polar(theta="y")

Controlando cores com ā€˜scalesā€™

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=dep_time))

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=dep_time)) +
  scale_color_gradient(low="#f7fcfd", high="#238b45" )

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=origin)) +
  scale_color_brewer(palette="Set2")

flights %>% group_by(origin, month) %>% 
  summarize(dep_delay_media=mean(dep_delay,na.rm=T)) %>% 
  ggplot() + 
  geom_tile(aes(x = origin, y = month, fill=dep_delay_media)) +
  scale_fill_gradient(low="#f7fcfd", high="#238b45")

flights %>% group_by(origin, month, carrier) %>% 
  tally() %>% 
  group_by(origin, month) %>%
  filter(n==max(n)) %>% 
  ggplot() + 
  geom_tile(aes(x = origin, y = month, fill=carrier)) +
  scale_fill_brewer(palette="Set2")

PersonalizaĆ§Ć£o de GrĆ”ficos alĆ©m de geometria

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=origin)) +
  ggtitle("RelaĆ§Ć£o entre distĆ¢ncia e duraĆ§Ć£o de voo, por aeroporto do Nova Iorque")

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=origin)) +
  ggtitle("RelaĆ§Ć£o entre distĆ¢ncia e duraĆ§Ć£o de voo, por aeroporto do Nova Iorque") +
  xlab("DistĆ¢ncia") +
  ylab("DuraĆ§Ć£o")

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=origin)) +
  ggtitle("RelaĆ§Ć£o entre distĆ¢ncia e duraĆ§Ć£o de voo, por aeroporto do Nova Iorque") +
  xlab("DistĆ¢ncia") +
  ylab("DuraĆ§Ć£o") + 
  theme(legend.position="bottom")

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=origin)) +
  ggtitle("RelaĆ§Ć£o entre distĆ¢ncia e duraĆ§Ć£o de voo, por aeroporto do Nova Iorque") +
  xlab("DistĆ¢ncia") +
  ylab("DuraĆ§Ć£o") +
    theme(axis.text.x = element_text(size=4),
          axis.text.y = element_text(size=4),
          axis.title.x = element_text(size=4),
          axis.title.y = element_text(size=4))

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=origin)) +
  ggtitle("RelaĆ§Ć£o entre distĆ¢ncia e duraĆ§Ć£o de voo, por aeroporto do Nova Iorque") +
  xlab("DistĆ¢ncia") +
  ylab("DuraĆ§Ć£o") +
  theme_classic()

# install.packages("ggthemes")
library(ggthemes)

flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=origin)) +
  ggtitle("RelaĆ§Ć£o entre distĆ¢ncia e duraĆ§Ć£o de voo, por aeroporto do Nova Iorque") +
  xlab("DistĆ¢ncia") +
  ylab("DuraĆ§Ć£o") +
  theme_economist()

flights %>% ggplot() +
  geom_bar(aes(x=origin))

ggsave("flights_barplot.png")

GrƔficos interativas e animaƧƵes

library(plotly)

graf_1 <- flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=origin)) +
  ggtitle("RelaĆ§Ć£o entre distĆ¢ncia e duraĆ§Ć£o de cada voo, por aeroporto do Nova Iorque em 2013") +
  xlab("DistĆ¢ncia") +
  ylab("DuraĆ§Ć£o") +
  theme_classic() 

graf_1 %>%
  ggplotly()
graf_2 <- flights %>% sample_n(1000) %>% ggplot() + 
  geom_point(aes(x = distance, y = air_time, color=origin, frame=month)) +
  ggtitle("RelaĆ§Ć£o entre distĆ¢ncia e duraĆ§Ć£o de cada voo, por aeroporto do Nova Iorque em 2013") +
  xlab("DistĆ¢ncia") +
  ylab("DuraĆ§Ć£o") +
  theme_classic() 

graf_2 %>%
  ggplotly()